We have studied the ability of cyclosporin A (CsA) and a non-immunosuppressive analogue, O-acetyl cyclosporin A (OACsA, B3-243) to inhibit the growth of human lung cancer cells in vitro. Using continuous drug exposure and the MTT colorimetric assay to determine cell growth we found that CsA produced partial growth inhibition at doses ranging from 0.5 to 3.0 micrograms ml-1 (0.4-2.4 microM). At progressively higher doses, complete growth inhibition and in situ cell lysis were seen. The P-glycoprotein expressing multidrug resistant (MDR) variant H69/LX4 of the small cell line H69/P was less sensitive to cyclosporins than the parent line, but this was not true of the non-P-glycoprotein expressing MDR variants of large cell line COR-L23 or adenocarcinoma line MOR. Sensitivity to OACsA was approximately 2-fold higher than that to CsA in most of the lines although not in the most sensitive line, COR-L88. Even in COR-L88, exposed to CsA or OACsA for 24 h, clonogenic cell survival was reduced only to 50%. There was no reduction in polyamine content of COR-L23 or COR-L88 cells following 48 h of exposure to CsA or OACsA. The effects on cell growth could not be inhibited by the addition of exogenous putrescine, nor could they be enhanced by the addition of alpha-difluoromethylorthinine. It does not appear therefore that inhibition of polyamine synthesis is the basis of the observed growth inhibition.

Mucosal ornithine decarboxylase activity and polyamine content has been proposed as a possible marker for malignant potential in gastrointestinal mucosa. Polyamine content and histological findings were examined in 107 pairs of endoscopic biopsy specimens taken from gastric fundus, fundic and specialised Barrett's oesophagus and Barrett's adenocarcinoma. The content of putrescine (median nmol/mg protein, range) the primary product of ornithine decarboxylase showed a progressive increase from gastric fundus (0.41, 0-15-1.5); fundic (0.45, 0.01-4.08); specialised Barrett's oesophagus (0 54, 0.01-2.0); dysplastic columnar lined oesophagus (0 56, 0.31-3.1) to adenocarcinoma (1.23, 0.29-8.98). Adenocarcinoma putrescine content was significantly greater than gastric fundus (p<0018) and fundic (p<003). Mucosal spermine, spermidine, and total polyamine values were greater in gastric fundus than fundic, specialised Barrett's oesophagus, and dysplastic columnar lined oesophagus (all p<0.001) suggesting failure to further metabolise putrescine to its higher polyamines in the metaplastic epithelium. Although metaplastic columnar lined oesophagus shows significant differences in polyamine metabolic activity from the stomach the important distinction between specialised and dysplastic columnar lined oesophagus cannot be made by measuring the polyamine content.

Gastric body mucosal proliferation was quantified and localised under conditions of increased gastrin drive using a variety of techniques. Rats were given omeprazole 400 ,imollkg/day by gavage and after 30 days mean serum gastrin rose 11-fold (p<0.001). Total mucosal polyamines rose 220% from 15.9 to 50.9 nmollmg protein (p<0001). This was associated with a 238% increase in crypt cell production rate from 0.541 to 1.83 crypt cells/h by vincristine metaphase arrest (p<0-02). Using computer aided counting of proliferating cell nuclear antigen (PCNA) immunostained nuclei to assess epithelial proliferation in hypergastrinaemia rat stomach: mucus neck cell PCNA labelling was increased by 41% (p<0.001) and gland cell PCNA labelling was increased by 222% (p<0.001). PCNA/AgNOR (argyrophilic nuclear organiser regions) co-stained sections were used to assess proliferative activity in cycling and noncycling cell populations. Data from these experiments suggest that, in addition to increasing the number of mucosal cells in cycle, cell life and cell cycle duration may be reduced in hypergastrinaemia.

Mammalian fibroblasts were cultured in the presence of alpha-methylornithine and/or methylglyoxal bis(guanylhydrazone), which inhibit the synthesis of polyamines. This led to a decrease in the cellular content of the polyamines spermine and spermidine by up to 60% when the cells were grown in the presence of both drugs together. The activity of the chromatin-associated enzyme ADP-ribosyltransferase was enhanced 2-3-fold in the drug-treated cells when measured in cells subsequently rendered permeable to exogenous NAD+, the substrate for the transferase. This is a novel and surprising observation, since the transferase is invariably activated by the addition of polyamines to a suitable incubation system such as permeabilized cells, isolated nuclei or the purified enzyme. We found no evidence that the activation was due to the appearance of DNA strand breaks, by using a variety of procedures including both neutral [the 'nucleoid' technique of Cook & Brazell [(1975) J. Cell Sci. 19, 261-279; (1976) J. Cell Sci. 22, 287-302]] and alkaline sucrose-gradient centrifugation and gel electrophoresis, suggesting that this therefore may not be the only means of regulating the activity of ADP-ribosyltransferase and that polyamines may have a role to play in this regard in vivo.

Summary We have examined the effects of Cyclosporin A (CsA) on growth and polyamine metabolism of , human T lymphoblastic leukaemia cells to ascertain the role of the polyamine biosynthetic pathway in the antitumour action of CsA. We observed that CsA had a dose-dependent inhibitory effect on growth of the cells in vitro, decreasing protein content, cell number and the rate of incorporation of 3H-thymidine into the cells. However, CsA treatment had no significant effect on intracellular polyamine levels in the cells. Contrary to previous reports, simultaneous addition of the diamine, putrescine, with CsA did not block or lessen the growth inhibitory effects of CsA. On the other hand, ornithine decarboxylase activity, the rate limiting enzyme of polyamine biosynthesis which converts ornithine to putrescine, was decreased by CsA treatment. This decrease appeared to be reversible and contrasts with the inhibition by x-difluoromethylornithine, which is irreversible and can be overcome by addition of putrescine. This suppression of ornithine decarboxylase by CsA is more likely to occur by indirect effects on translation and/or transcription rather than a direct effect on the enzyme. It may be a contributory factor in the overall antiproliferative effects of CsA but is more likely to be a response to these growth inhibitory effects rather than a direct effect of the drug.